thewall commited on
Commit
45a0e8c
·
1 Parent(s): ae7ddd5

Upload jolma_split.py

Browse files
Files changed (1) hide show
  1. jolma_split.py +159 -0
jolma_split.py ADDED
@@ -0,0 +1,159 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import re
2
+ import pandas as pd
3
+ import datasets
4
+ from functools import cached_property, cache
5
+
6
+ logger = datasets.logging.get_logger(__name__)
7
+
8
+
9
+ _CITATION = """\
10
+ @article{jolma2010multiplexed,
11
+ title={Multiplexed massively parallel SELEX for characterization of human transcription factor binding specificities},
12
+ author={Jolma, Arttu and Kivioja, Teemu and Toivonen, Jarkko and Cheng, Lu and Wei, Gonghong and Enge, Martin and \
13
+ Taipale, Mikko and Vaquerizas, Juan M and Yan, Jian and Sillanp{\"a}{\"a}, Mikko J and others},
14
+ journal={Genome research},
15
+ volume={20},
16
+ number={6},
17
+ pages={861--873},
18
+ year={2010},
19
+ publisher={Cold Spring Harbor Lab}
20
+ }
21
+ """
22
+
23
+ _DESCRIPTION = """\
24
+ PRJEB3289
25
+ https://www.ebi.ac.uk/ena/browser/view/PRJEB3289
26
+ Data that has been generated by HT-SELEX experiments (see Jolma et al. 2010. PMID: 20378718 for description of method) \
27
+ that has been now used to generate transcription factor binding specificity models for most of the high confidence \
28
+ human transcription factors. Sequence data is composed of reads generated with Illumina Genome Analyzer IIX and \
29
+ HiSeq2000 instruments. Samples are composed of single read sequencing of synthetic DNA fragments with a fixed length \
30
+ randomized region or samples derived from such a initial library by selection with a sequence specific DNA binding \
31
+ protein. Originally multiple samples with different "barcode" tag sequences were run on the same Illumina sequencing \
32
+ lane but the released files have been already de-multiplexed, and the constant regions and "barcodes" of each sequence \
33
+ have been cut out of the sequencing reads to facilitate the use of data. Some of the files are composed of reads from \
34
+ multiple different sequencing lanes and due to this each of the names of the individual reads have been edited to show \
35
+ the flowcell and lane that was used to generate it. Barcodes and oligonucleotide designs are indicated in the names of \
36
+ individual entries. Depending of the selection ligand design, the sequences in each of these fastq-files are either \
37
+ 14, 20, 30 or 40 bases long and had different flanking regions in both sides of the sequence. Each run entry is named \
38
+ in either of the following ways: Example 1) "BCL6B_DBD_AC_TGCGGG20NGA_1", where name is composed of following fields \
39
+ ProteinName_CloneType_Batch_BarcodeDesign_SelectionCycle. This experiment used barcode ligand TGCGGG20NGA, where both \
40
+ of the variable flanking constant regions are indicated as they were on the original sequence-reads. This ligand has \
41
+ been selected for one round of HT-SELEX using recombinant protein that contained the DNA binding domain of \
42
+ human transcription factor BCL6B. It also tells that the experiment was performed on batch of experiments named as "AC".\
43
+ Example 2) 0_TGCGGG20NGA_0 where name is composed of (zero)_BarcodeDesign_(zero) These sequences have been generated \
44
+ from sequencing of the initial non-selected pool. Same initial pools have been used in multiple experiments that were \
45
+ on different batches, thus for example this background sequence pool is the shared background for all of the following \
46
+ samples. BCL6B_DBD_AC_TGCGGG20NGA_1, ZNF784_full_AE_TGCGGG20NGA_3, DLX6_DBD_Y_TGCGGG20NGA_4 and MSX2_DBD_W_TGCGGG20NGA_2
47
+ """
48
+
49
+ _DOWNLODE_MANAGER = datasets.DownloadManager()
50
+ _RESOURCE_URL = "https://huggingface.co/datasets/thewall/DeepBindWeight/resolve/main"
51
+ SELEX_INFO_FILE = _DOWNLODE_MANAGER.download(f"{_RESOURCE_URL}/ERP001824-deepbind.xlsx")
52
+ PROTEIN_INFO_FILE = _DOWNLODE_MANAGER.download(f"{_RESOURCE_URL}/ERP001824-UniprotKB.xlsx")
53
+ pattern = re.compile("(\d+)")
54
+ URL = "https://huggingface.co/datasets/thewall/jolma_split/resolve/main"
55
+
56
+
57
+ class JolmaSplitConfig(datasets.BuilderConfig):
58
+ def __init__(self, protein_prefix="", protein_suffix="", max_length=1000, max_gene_num=1,
59
+ aptamer_prefix="", aptamer_suffix="", **kwargs):
60
+ super(JolmaSplitConfig, self).__init__(**kwargs)
61
+ self.data_dir = kwargs.get("data_dir")
62
+ self.protein_prefix = protein_prefix
63
+ self.protein_suffix = protein_suffix
64
+ self.aptamer_prefix = aptamer_prefix
65
+ self.aptamer_suffix = aptamer_suffix
66
+ self.max_length = max_length
67
+ self.max_gene_num = max_gene_num
68
+
69
+
70
+ class JolmaSubset(datasets.GeneratorBasedBuilder):
71
+
72
+ SELEX_INFO = pd.read_excel(SELEX_INFO_FILE, index_col=0)
73
+ PROTEIN_INFO = pd.read_excel(PROTEIN_INFO_FILE, index_col=0)
74
+
75
+ BUILDER_CONFIGS = [
76
+ JolmaSplitConfig(name=key) for key in ["p70c10s61087t100", "p99c3s325414t1000"]
77
+ ]
78
+
79
+ DEFAULT_CONFIG_NAME = "p70c10s61087t100"
80
+
81
+ def _info(self):
82
+ return datasets.DatasetInfo(
83
+ description=_DESCRIPTION,
84
+ features=datasets.Features(
85
+ {
86
+ "id": datasets.Value("int32"),
87
+ "identifier": datasets.Value("string"),
88
+ "seq": datasets.Value("string"),
89
+ "quality": datasets.Value("string"),
90
+ "count": datasets.Value("int32"),
91
+ "protein": datasets.Value("string"),
92
+ "protein_id": datasets.Value("string"),
93
+ }
94
+ ),
95
+ homepage="https://www.ebi.ac.uk/ena/browser/view/PRJEB3289",
96
+ citation=_CITATION,
97
+ )
98
+
99
+ @cached_property
100
+ def selex_info(self):
101
+ return self.SELEX_INFO.loc[self.config.name]
102
+
103
+ @cached_property
104
+ def protein_info(self):
105
+ return self.PROTEIN_INFO.loc[self.config.name]
106
+
107
+ def design_length(self):
108
+ return int(pattern.search(self.protein_info["Ligand"]).group(0))
109
+
110
+ def get_selex_info(self, sra_id):
111
+ return self.SELEX_INFO.loc[sra_id]
112
+
113
+ def get_protein_info(self, sra_id):
114
+ return self.PROTEIN_INFO.loc[sra_id]
115
+
116
+ @cache
117
+ def get_design_length(self, sra_id):
118
+ return int(pattern.search(self.get_protein_info(sra_id)["Ligand"]).group(0))
119
+
120
+ def _split_generators(self, dl_manager):
121
+ train_file = dl_manager.download(f"{URL}/{self.config.name}_train.csv.gz")
122
+ test_file = dl_manager.download(f"{URL}/{self.config.name}_test.csv.gz")
123
+ return [
124
+ datasets.SplitGenerator(name=datasets.Split.TRAIN, gen_kwargs={"filepath": train_file},),
125
+ datasets.SplitGenerator(name=datasets.Split.TEST, gen_kwargs={"filepath": test_file}, ),
126
+ ]
127
+
128
+ def _generate_examples(self, filepath):
129
+ """This function returns the examples in the raw (text) form."""
130
+ logger.info("generating examples from = %s", filepath)
131
+ data = pd.read_csv(filepath)
132
+ for key, row in data.iterrows():
133
+ sra_id = row["identifier"].split(":")[0]
134
+ protein_info = self.get_protein_info(sra_id)
135
+ proteins = protein_info["Sequence"]
136
+ gene_num = protein_info["Unique Gene"]
137
+ protein_id = protein_info["Entry"]
138
+ protein_seq = f"{self.config.protein_prefix}{proteins}{self.config.protein_suffix}"
139
+ aptamer_seq = f'{self.config.aptamer_prefix}{row["seq"]}{self.config.aptamer_suffix}'
140
+ if len(protein_seq)>self.config.max_length:
141
+ continue
142
+ if gene_num>self.config.max_gene_num:
143
+ continue
144
+ if str(proteins)=="nan" or len(str(proteins))==0:
145
+ continue
146
+ ans = {"id": key,
147
+ "protein": protein_seq,
148
+ "protein_id": protein_id,
149
+ "seq": aptamer_seq,
150
+ "identifier": row["identifier"],
151
+ "count": int(row["count"]),
152
+ "quality": row['quality']}
153
+ yield key, ans
154
+
155
+
156
+ if __name__=="__main__":
157
+ from datasets import load_dataset
158
+ dataset = load_dataset("jolma_split.py", split="all")
159
+